NRPB-W3: Non-technical summary

Non-technical summary

The Atmospheric Dispersion Modelling Liaison Committee consists of representatives of government departments, government agencies and organisations in the nuclear industry. It was established mainly to review current understanding of atmospheric dispersion and to identify suitable models for application primarily in the authorisation or licensing processes for the nuclear industry. The Committee places contracts to carry out pieces of work on topics of interest. This report presents the work done during the 2000-01 financial year.

Atmospheric dispersion calculations require input data describing the range of atmospheric conditions that might occur during the period over which the release is considered, and the relative frequency of different conditions. Ideally these data should be taken from measurements made at the site of the release. Unfortunately, many industrial sites do not take sufficiently detailed measurements of atmospheric conditions to provide the input required by modern dispersion models, and so data must be obtained from other sources. Measurements taken at a number of Meteorological Office sites for use in weather forecasting can be used for dispersion modelling. The number of suitable sites has reduced in recent years, but records exist from measurements taken in the past. The Committee placed a contract to review the most appropriate source of data for dispersion calculations. This considered older measurements from a meteorological site close to the release point, or more recent measurements from a more distant site, and the contractor concluded that more recent measurements are likely to give better modelling results. The study also suggested that suitable information could be obtained from some of the quantities used in weather forecasting.

The Committee also arranged a series of presentations on techniques for undertaking uncertainty analyses. Any model requires values to be given for a number of quantities used in the model, and there is generally some uncertainty in the most appropriate value to use for a particular application. The presentations covered ways of describing the possible range of values for the various quantities and ways of using that information to calculate the range of possible values for the model prediction.

Last reviewed: 1 September 2009